Download High-throughput sequencing identifies distinct fecal and

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Trimeric autotransporter adhesin wikipedia , lookup

Probiotic wikipedia , lookup

Horizontal gene transfer wikipedia , lookup

Bacteria wikipedia , lookup

Probiotics in children wikipedia , lookup

Marine microorganism wikipedia , lookup

Skin flora wikipedia , lookup

Bacterial cell structure wikipedia , lookup

Phospholipid-derived fatty acids wikipedia , lookup

Triclocarban wikipedia , lookup

Gut flora wikipedia , lookup

Metagenomics wikipedia , lookup

Bacterial morphological plasticity wikipedia , lookup

Community fingerprinting wikipedia , lookup

Bacterial taxonomy wikipedia , lookup

Human microbiota wikipedia , lookup

Transcript
High-throughput sequencing identifies distinct fecal and
mucosal gut microbiota correlating with different mucosal
proteins
Li-na Dong
1
2
3
1
, Jun-ping Wang Corresp.,
2
, Ping Liu
3
, Yun-feng Yang
2
, Jing Feng
2
, Yi Han
2
Central Laboratory, Shanxi Provincial People’s Hospital, Affiliate of Shanxi Medical University, Taiyuan, China
Department of Gastroenterology, Shanxi Provincial People’s Hospital, Affiliate of Shanxi Medical University, Taiyuan, China
Department of Gynaecology, Shanxi Provincial People’s Hospital, Affiliate of Shanxi Medical University, Taiyuan, China
Corresponding Author: Jun-ping Wang
Email address: [email protected]
The intestinal microbiota is associated with human health. The luminal microbiota (LM) and
mucosa-associated microbiota (MAM) are distinct ecosystems with different metabolic and
immunological functions. Several studies have examined the correlations between the gut
microbiota and clinical indices, but few have investigated the relationships between the
microbiota and mucosal proteins. We characterized the intestinal LM and MAM in Chinese
people and examined the association between these communities and the expression of
mucosal proteins. Fresh fecal samples and distal colonic mucosal biopsies were collected
from 32 subjects before (fecal) and during (mucosal) flexible sigmoidoscopy. We used
high-throughput sequencing targeting the 16SrRNA gene V3–V4 region to analyze the
samples and reverse transcription(RT)–PCR to detect the expression of colonic proteins
BDNF, ZO1, TLR2, TLR4, AQP3, and AQP8. Differences in the stool and mucosal microbiota
were identified and a correlation network analysis performed. The LM and MAM
populations differed significantly. In LM, the microbiota composition correlated significantly
positively with host age, and Firmicutes (phylum) correlated positively with body mass
index (BMI), but inversely with ZO1.At the genus level, systemic indices, such as age, BMI,
and BDNF, correlated predominantly with LM, whereas systemic and local indices, such as
TLR2, correlated with both MAM and LM. ZO1 and TLR4 which usually exert a local effect,
mainly correlated with MAM. Different bacteria were associated with the expression of
different proteins. Our data suggest that The microbial compositions of LM and MAM
differed. Different gut bacteria may play different roles by regulating the expression of
different proteins.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
1
High-throughput Sequencing Identifies Distinct Fecal and Mucosal Gut Microbiota
2
Correlating with Different Mucosal Proteins
3
Li-na Dong et al. Different Fecal and Mucosal Microbiota.
4
Li-na Dong1, Jun-ping Wang2*, Ping Liu3, Yun-feng Yang2, Jing Feng2, Yi Han2
5
*Corresponding author:
6
Jun-ping Wang. Tel: +86-0351-4960141;E-mail: [email protected]
Name
Highest academic degree
Email
Postal addresses
Li-na Dong
MSc
[email protected]
Shanxi Provincial
Jun-ping
PhD
[email protected]
People’s
Hospital, 29
Wang
Shuangta Road,
Ping Liu
MSc
[email protected]
Yun-feng
MSc
[email protected]
Jing Feng
MSc
[email protected]
Yi Han
MSc
[email protected]
Taiyuan 030012,
China
Yang
7
1Central
8
Taiyuan, China.
Laboratory, Shanxi Provincial People’s Hospital, Affiliate of Shanxi Medical University,
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
9
2Department
of Gastroenterology, Shanxi Provincial People’s Hospital, Affiliate of Shanxi
10
Medical University, Taiyuan, China
11
3Department
12
University, Taiyuan, China
13
Author contribution:Li-na Dong performed the majority of experiments, analyzed the data and
14
drafted the manuscript ; Jun-ping Wang, Ping Liu participated in the design of this study and
15
manuscript reviewer, Yun-feng Yang, Jing Feng, Yi Han performed specimen collection. All
16
authors read and approved the final manuscript.
17
Supported by the International Science and Technology Cooperation Project of Shanxi
18
(No.2013081066), Science Foundation of Health and family planning commission of Shanxi
19
Province(No.201201059).The funding bodies had no role in the study design, data collection and
20
analysis, decision to publish, or preparation of the manuscript.
21
Institutional review board statement: The study was approved by the Ethics Committee of
22
Shanxi People's Hospital.
23
Conflict-of-interest statement:The authors declare that they have no competing interests.
24
Data sharing statement: No additional data are available.
25
Correspondenceto:Wang Jun-ping, Professor, Department of Gastroenterology, Shanxi
26
Provincial People’s Hospital, Affiliate of Shanxi Medical University, Shanxi Provincial People,s
27
Hospital, 29 Shuangta Road, Taiyuan 030012, China.
28
E-mail:[email protected]
29
Tel: +86-0351-4960141
of Gynaecology, Shanxi Provincial People’s Hospital, Affiliate of Shanxi Medical
30
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
31
ABSTRACT
32
The intestinal microbiota is associated with human health. The luminal microbiota (LM) and
33
mucosa-associated microbiota (MAM) are distinct ecosystems with different metabolic and
34
immunological functions. Several studies have examined the correlations between the gut
35
microbiota and clinical indices, but few have investigated the relationships between the
36
microbiota and mucosal proteins.We characterized the intestinal LM and MAM in Chinese
37
people and examined the association between these communities and the expression of mucosal
38
proteins. Fresh fecal samples and distal colonic mucosal biopsies were collected from 32 subjects
39
before (fecal) and during (mucosal) flexible sigmoidoscopy. We used high-throughput
40
sequencing targeting the 16SrRNA gene V3–V4 region to analyze the samples and reverse
41
transcription(RT)–PCR to detect the expression of colonic proteins BDNF, ZO1, TLR2, TLR4,
42
AQP3, and AQP8. Differences in the stool and mucosal microbiota were identified and a
43
correlation network analysis performed. The LM and MAM populations differed significantly.
44
In LM, the microbiota composition correlated significantly positively with host age, and
45
Firmicutes (phylum) correlated positively with body mass index (BMI), but inversely with
46
ZO1.At the genus level, systemic indices, such as age, BMI, and BDNF, correlated
47
predominantly with LM, whereas systemic and local indices, such as TLR2, correlated with both
48
MAM and LM. ZO1 and TLR4 which usually exert a local effect, mainly correlated with MAM.
49
Different bacteria were associated with the expression of different proteins.
50
that The microbial compositions of LM and MAM differed. Different gut bacteria may play
51
different roles by regulating the expression of different proteins.
Our data suggest
52
53
Keywords Gastrointestinal microbiota·high-throughput sequencing·16S rRNA gene Mucosal
54
Proteins
55
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
56
57
58
Introduction
59
The intestinal microbiota is a complex community of Bacteria, Archaea, viruses, and Eukarya. A
60
wide variety of bacterial species in the gastrointestinal tract exert numerous effects on the host
61
and influence a variety of gastrointestinal functions[1]. Fecal samples (representing the luminal
62
niche) are examined in most studies ofthe intestinal microbiota because they are easily collected.
63
However, recent researchhas shown that the microbial compositions ofthe luminal microbiota
64
(LM) and the mucosa-associated microbiota (MAM) differ, suggesting that these two distinct
65
microbial populations play different roles within the intestinal microbiota ecosystem [2]. LM is
66
the microbiota involved in the whole intestine, whereas MAM represents a special niche.
67
Because MAM is in close contact with the host, it may play a more prominent role in the
68
intestinal barrier function, whereas LM may play a key role in metabolic activities and nutrient
69
harvest[3]. However, the different functions ofLM and MAM are unknown.
70
71
Many proteins are involved in the colonic microbe–host interactions. The tight junction proteins
72
constitute a critical platform that regulates the integrity of the epithelial barrier and maintains the
73
activation of the mucosal immunitywithin an acceptable range [4].The effects of brain-derived
74
neurotrophic factor (BDNF) in the gut are beginning to be identified; there is growing evidence
75
that BDNF also plays an important role in gastrointestinal functions. Wang et al. found that the
76
activation of PAR-2 signaling by fecal supernatants from patients withirritable bowel syndrome
77
(IBS) with diarrhea promoted the expression of colonic BDNF, thereby contributing to IBS-like
78
visceral hypersensitivity [5]. Toll-like receptors (TLRs) are pattern recognition receptors
79
expressed by various cells in the gastrointestinal tract. The microbiota may directly interact with
80
the TLRs and regulate the gut immune responses, especially through the activation of TLRs [6].
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
81
Water transport through the human digestive system is physiologically crucial for maintaining
82
the water homeostasis of the body and ensuring digestive and absorptive equilibria. Aquaporins
83
(AQPs) are important transmembrane water channel proteins. Guttman et al. foundthat the
84
altered localization of AQPs was partly dependent on the bacterial type III effector proteins EspF
85
and EspG[7].
86
87
We speculate that different bacteria play different roles in the colon, and that different bacteria
88
regulate the expression of different proteins,thus affecting intestinal functions. However, few
89
data are available on the correlation between the intestinal microbiota and mucosa-associated
90
proteins.
91
92
In this study, we used high-throughput pyrosequencing of the bacterial 16S rRNA gene to
93
compare the microbial communities in the fecesand mucosa of Chinese subjects, and to study
94
their association with the expression of colonic mucosal proteins (ZO1, BDNF, TLR2, TLR4,
95
AQP3, and AQP8) and the clinical features (age and body mass index [BMI]) of the host.
96
97
Materials and Methods
98
99
Study Subjects
100
Thirty-two Chinese patients were recruited from the Department of Gastroenterology, Shanxi
101
Provincial People’s Hospital,in 2013 and 2014. None of the subjects enrolled in the study had
102
taken corticosteroids, opioids, or antibiotics in the 6 months preceding the study; none had any
103
systemic comorbidity;andnone had a history of excessive alcohol intake (>20 alcoholic drinks
104
per week). Patients with a prior history of gastrointestinal surgery or intestinal organic disease
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
105
were excluded. All subjects gave their signed informed consent before participation. The study
106
was performed in accordance with the principles of the Declaration of Helsinki, and the study
107
protocol was approved by the Ethics Committee of Shanxi Provincial People’s Hospital, China.
108
109
Stool Sample Processing and DNA Extraction
110
The fecal samples were collected at home <12 h before colonoscopy, frozen immediately at-20
111
°C, and transported within 12 h to the study center, where they were stored at-80 °C until
112
analysis. Bacterial DNA was extracted from the fecal samples withthe QIAamp® DNA Stool
113
Mini Kit (Qiagen, Hilden, Germany), according to the manufacturer’s protocol. The DNA
114
concentrations were quantified with an Eppendorf BioSpectrometer® (Eppendorf,Hamburg,
115
Germany).
116
117
Genomic DNA Extraction
118
Total genomic DNA was extracted from samples of digests from the colon with a QIAamp DNA
119
Mini Kit (Qiagen), according to the manufacturer’s instructions. The concentration and purity of
120
the genomic DNA were measured with an Eppendorf BioSpectrometer.
121
122
PCR Amplification of V3–V4 Region of Bacterial 16S rRNA Gene and Illumina Sequencing
123
The bacterial genomic DNA was used as the template to amplify the V3–V4 hypervariable
124
region of the 16SrRNA gene with the forward primer (5-GACTACHVGGGTATCTAATCC-
125
3)and the reverse primer (5-CCTACGGGNGGCWGCAG-3).
126
127
Bioinformatic Analysis
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
128
Pairs of reads from the original DNA fragments were merged using FLASH, which was designed
129
to merge pairs of reads when the original DNA fragments were shorter than twice the read length.
130
The sequencing reads were assigned to each sample according to a unique barcode and were
131
analyzed with the QIIME (Quantitative Insights Into Microbial Ecology) software package and
132
the UPARSE pipeline. In brief, the reads were filtered with the QIIME quality filters using the
133
default settings for Illumina processing, and the operational taxonomic units (OTUs) were
134
selected using the UPARSE pipeline. The samples were sequenced on an Illumina MiSeq
135
Benchtop Sequencer and the bioinformatic analysis were performed by Genesky Biotechnologies
136
Inc., Shanghai, China.
137
The sizes of the bacterial groups were expressed as percentages of the total bacteria.
138
139
Quantitative Real-time Polymerase Chain Reaction (qPCR)
140
The total mucosal RNAs were extracted from the colonic biopsies using the TaKaRa MiniBEST
141
Universal RNA Extraction Kit (TaKaRa), according to the manufacturer’s instructions. The
142
mRNA concentrations and purity were measured with an Eppendorf BioSpectrometer. After
143
reverse transcription with PrimeScript Reverse Transcriptase Mix (TaKaRa), which converted
144
the total RNA to cDNA, the expression of the BNNF, ZO1, TLR2, TLR4, AQP3, and AQP8genes
145
was determined with qPCR and SYBR Green technology on a Bio-Rad CFX96™ Q-PCR
146
instrument(Bio-Rad, USA) for each sample in duplicate. The specific primers are listed in
147
Table1. Each amplification reaction was run in duplicate in a final volume of 20l containing
148
10l of Power SYBR Green PCR master mix, 400 nmol of the forward and reverse primers, and
149
1l of cDNA. All the qPCRs were optimized and performed in 0.2ml 96-well plates,with the
150
following cycling program: initial denaturation at 95 °C for 10 min, followed by 40 cycles of 95
151
°C for 15 s and 60 °C for 30s. Fluorescence was measured at the last step of each cycle. To
152
determine the specificity of the amplification, the dissociation characteristics of the double-
153
stranded DNA were determined with a melting curve analysis. The dissociation of the PCR
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
154
products was monitored by slowly heating them,in increments of 0.1 °C/s, from 55 °C to 95 °C,
155
with fluorescence measurements made at 0.1 °C intervals. The correct PCR product length was
156
confirmed with gel electrophoresis. Negative controls lacking the template DNA were included
157
in triplicate. Standard curves for the target bacterial groups were generated using serial dilutions
158
(corresponding to approximately 101–1010copies/l) of the purified and quantified PCR products
159
generated from genomic DNA with standard PCR. The mRNA levelswere absolutely quantified
160
by converting the sample cycle threshold (Cq) values to concentrations (copies per l) based on
161
the standard curves [8].
162
163
Statistical Analysis
164
All statistical analyses were performed with SPSS 22.0 for Windows (SPSS Inc., USA).To
165
determine the statistical differences between the two groups, we usedan independent-
166
samplesttest and the Mann–Whitney test. Correlations were determined with Spearman’s
167
correlation. The resulting pvalues were adjusted using the Benjamini–Hochberg false discovery
168
rate (FDR) correction. Only FDR-corrected p values below 0.05 were considered significant.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
170
Results
171
Study Population
172
We investigated 64 samples from 32 subjects. All subjects provided both a fecal sample and a
173
colonic mucosal sample. The study population consisted of 50% females,and had a mean age of
174
49 (20–65) years and a mean BMI of 23.24.
175
Characteristics of the Pyrosequencing Results
176
We obtained a total of 4,351,929 raw reads and 3,545,053 reads remained after filtering. The
177
sequencing analysis of the 64 samples identified 1026 OTUs. The rarefaction curves tended to
178
approach the saturation plateau, indicating that the number of samples used in this study was
179
reasonable. The same tendency was found in the Shannon–Wiener curves, indicating that the
180
database of 16S rRNA gene sequences was very abundant and reflected the vast majority of
181
microbial information.
182
Microbial Population Structures in theIntestinal Lumen and Mucosa
183
The gut fecal samples showed significantly more diversity and richness than the mucosal
184
samples (Table 2), and LM and MAM differed significantly (Fig.1A and B).
185
Significant differences between LM and MAM were identified in almost all populated phyla.
186
Bacteroidetes (44.7%) andFirmicutes(42.2%) were the most strongly represented phyla in
187
LM,followed by Proteobacteria(8.5%), whereas Proteobacteria (56.6%) was the most strongly
188
represented
189
Bacteroidetes(12.7%)(P<0.05;Figure 2).
190
At the genus level, the relative abundances of Escherichia–Shigella, Streptococcus, Clostridium
191
sensu stricto1,Sphingomonas, Acinetobacter, Brevundimonas, and Enhydrobacter were
192
significantly greater in MAM than in LM,whereas those of Bacteroides, Faecalibacterium,
193
incertae
phylum
sedis,
in
MAM,
Subdoligranulum,
followed
Pseudobutyrivibrio,
by
Firmicutes(20.2%)
Megasphaera,
and
Parasutterella,
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
194
Akkermansia, Alistipes, and Lachnospira were significantly lower in MAM than in LM (P <
195
0.05).
196
Correlation with Age and BMI
197
Correlations with age and BMI were only detected in LM,and not in MAM. Fecal microbial
198
diversity correlated significant positively with the age of the host (r=0.34, P=0.05). In LM, the
199
proportions of phylum Firmicutes, class Clostridia(r = 0.398, P= 0.024), order Clostridiales (r =
200
0.398, P= 0.024), and family Ruminococcacea (r = 0.359, P= 0.043) correlated positively with
201
age, whereas the proportion of family Bacteroides (r = –0.437, P= 0.012) correlated negatively
202
with age.
203
204
In LM, class Bacteroidia(r= –0.367, P=0.039) and LM order Bacteroidales (r= –0.367,P=0.039)
205
correlated negatively with BMI. In the fecal microbiota (LM),the proportions of phylum
206
Firmicutes (r=0.480, P=0.018<0.05),class Coriobacteriia in phylum Actinobacteria(r=0.528,
207
P=0.002), order Coriobacteria (r=0.504,P=0.007), family Coriobacteriaceae(r=0.504, P=0.007),
208
genus Collinsella (r=0.435, P=0.013), and class Chloroplast inphylum Cyanobacteria(r=0.433,
209
P=0.013) correlated positively with BMI.
210
211
Correlation with Mucosal Proteins BDNF, ZO1, TLR2, TLR4, AQP3, and AQP8
212
In a correlation analysis of bacterial abundance and the expression of mucosal proteins, distinct
213
gut microbiota correlated with the expression ofBDNF (Table3), ZO1(Table4), TLR2(Table5),
214
and TLR4(Table6). The bacteria that correlated with protein expression, age, and BMI mainly
215
belonged to the phylum Firmicutes. The bacteria that correlated with BDNF expression, age, and
216
BMI belonged to LM, whereas the bacteria that correlated with TLR4 and ZO1 expression
217
belonged to MAM. Bacteria belonging to both LM and MAM correlated with TLR2 expression,
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
218
and the trends in LM and MAM were consistent. Correlationswere detected with AQP8,but not
219
with AQP3. In LM, the phylum Cyanobacteria, phylum Bacteroidetes, and genus Prevotella
220
correlated negatively with AQP8,and in MAM, the phylum Firmicutesand genus Clostridiales
221
correlated negatively with AQP8. However, in MAM, the phylum Proteobacteriaand order
222
Caulobacterales correlated positively with AQP8.
223
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
224
Discussion
225
Different habitats in the human body harbor distinct microbiota,which can be divided into
226
different groups according to their anatomicallocations [9]. Mucosal samples were obtained from
227
the large bowel before and after its preparation for an endoscopic procedure. Whether this bowel
228
preparation affects the mucosal microbiota is still controversial. To avoid any interference, we
229
collected the mucosal samples after bowel preparation. We used deep sequencing to
230
determinethe bacterial compositions of the microbiota in the fecal samples and mucosal samples.
231
More than 95% of the sequences in all the stool and mucosalsamples belonged to the three most
232
popular bacterial phyla, Firmicutes, Bacteroidetes, and Proteobacteria. This isconsistent with the
233
findings of previous studies, which showed that these phyla account for the majority of the gut
234
microbiota
235
significantlygreaterbacterial diversity and richness than the mucosal samples,as in the study of
236
Ringel et al. [2]. Comparing the proportions of the dominant bacterial taxa in the fecal and
237
mucosal samples revealed significant differences. In this study,Proteobacteria was the
238
predominant phylum in MAM.This differs from other reports, perhaps reflecting geographic
239
differences, because it is well known that the Chinese diet and genetics are very differentfrom
240
those in western countries, and these factors markedly influence the gut microbiota. Furthermore,
241
MAM was sampled from a unique location, whereas LM was sampled from the whole intestinal
242
microbiota, so MAM may have a more specific relationship with the host.
243
We performed a correlation analysis ofthe two bacterialpopulationsat the phylum, class, order,
244
family, and genus levels,withsix proteins, host age, and host BMI. The results showed that
245
although the Proteobacteria was the predominant phylum in MAM, the bacteria that correlated
246
with specific proteins, age, and BMI mainly belonged to the phylum Firmicutes. In accordance
247
with their distinct microbial compositions, LM and MAM showed different correlations. The
248
bacteria that correlated with BDNF, age, and BMI belonged to LM.In contrast, the bacteria that
249
correlated with TLR4 and ZO1 belonged to MAM. Bacteria that correlated with TLR2 belonged
250
to both LM and MAM, and the trend was consistent in LM and MAM.
in
both
stool
and
mucosal
samples.
The
fecal
samples
displayed
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
251
Age and BMI are important factors influencing the composition of the microbiota. Fecal
252
microbial diversity correlated significantly positively with age,suggesting that microbial
253
diversity increases with age. Members of the phylaBacteroidetes and Firmicutes were the main
254
kinds of bacteria the microbiota, but they correlated oppositely with age and BMI. Based on the
255
sequencing results, in LM,the proportions of the family Ruminococcaceae(belonging to
256
classClostridia) andthe phylum Firmicutes correlated positively with age, whereasthe proportion
257
of the family Bacteroides(belonging to phylum Bacteroidetes) correlated negatively with age.
258
The phylum Bacteroides benefitshuman health, and these bacteria are reduced in older
259
people[10]. It has been suggested that bacterial communities also undergo an aging process. In
260
the LM microbiota, the phylum Firmicutes correlated positively with BMI,whereasthe phylum
261
Bacteroidetes correlated negatively with BMI. Recent research has identified relationships
262
between the bacterial composition of the gut microbiota and obesity. However, the results of
263
studies of obesity and phylum-level changes in the gut microbiota are frequently
264
contradictory,whichrequires
265
methodological issues or geographic factors [11]. Interestingly, the phylumFirmicutes and class
266
Coriobacteriia(belonging to the phylumActinobacteria) had opposite relationships with BMI and
267
ZO1.As the numbers ofFirmicutes and Coriobacteriiaincreased, BMI increased, whereas the
268
expression of ZO1 decreased. Many studies have shown that some bacteria are associated with
269
BMI, and BMIsoutside the normal range are related to many diseases. Our results suggest that
270
the interaction between bacteria and BMI maybe related to the expression of ZO1.
271
Like age and BMI, the bacteria that correlated with BDNF expression belonged to LM, but these
272
bacteria only belonged to the phylum Firmicutes. Interestingly, all the bacteria that correlated
273
with BDNF expression also correlated with TLR2 expression, and with the same trends. As the
274
bacteria in the genus Faecallibacterium(family Ruminococcaceae, order Clostridiales, class
275
Clostridia), the genus Lachnospira (familyLachnospiraceae,order Clostridiales, class Clostridia),
276
and the order Lactobacillales(class Bacilli)increased, the expression of BDNF and TLR2
277
decreased. BDNF also correlated positively with TLR2. Although positive correlations also
explanation.
These
discrepancies
may
be
attributable
to
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
278
existed between BDNF and TLR4 and between TLR2 and TLR4, the bacteria associated with
279
TLR4 expression differed greatly from those related tothe expression of BDNF and TLR2. The
280
bacteria that correlated with TLR4 were only found in MAM. Some bacteria that negatively
281
correlated with TLR2 were also found in MAM, but they only belonged to the class Bacilli,
282
whereas TLR4 expression was mainly associated with the class Clostridia. The enteric
283
commensal bacteria of the genusFaecallibacterium,which belongs to the Clostridium group,exert
284
an anti-inflammatory effect.In the present study, Faecallibacterium correlated negatively with
285
the expression of BDNF, TLR2, and TLR4. Similarly, the order Lactobacillales(class Bacilli)also
286
correlated negatively with BDNF, TLR2, and TLR4. Lactobacillales and Faecallibacterium may
287
display similar effects because Lactobacillales can cause the numbers ofFaecallibacteriumto
288
increase and the TLRs may exert an anti-inflammatory effect.
289
Although the phylumProteobacteriawas the most strongly represented bacterial taxon in MAM,
290
only a few bacteria from the Proteobacteria correlated with the expression of specific proteins:
291
MAM members of the class Betaproteobacteria correlated positively with ZO1 and MAM
292
members of the family Mitochndric(order Rickettsiales, class Alphaproteobacteria) correlated
293
positively with TLR2. The genusHaemophilus(family Pasteurellaceae, order Pasteurellales, class
294
Gammaproteobacteria)contains common neutral or pathogenic bacteria, andin MAM,
295
genusHaemophilus
296
LM,Haemophilus also correlated negatively with TLR2. Round et al. showed that unlike
297
pathogens whose TLR ligands trigger inflammation, some commensal bacteria exploit the TLR
298
pathway to actively suppress immune reactions [12]. Our findings indicate that commensal
299
microbes avoid activating TLR and that the pathogenicity of Haemophilusmay be related to the
300
inhibition of the primary immune response.
301
ZO1 is an important tight junction protein.Six MAM bacteria correlated positively with ZO1
302
expression. Therefore, increases in these bacteria may increase the levels of ZO1, partly restoring
303
the function of the intestinal barrier. AQPs areessential proteins in water metabolism, and the
304
different AQP proteins are expressed in different locations, with AQP3 and AQP8 present in the
correlated
negatively
with
BDNF,
TLR2,
and
TLR4,
and
in
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
305
colon. In this study,bacteria correlated with AQP8,but not with AQP3. Fecal Cyanobacteria
306
(phylum) correlated negatively with AQP8. In LM, the genus Prevotella genus (phylum
307
Bacteroidetes) correlated negatively with AQP8, as did the genusClostridium(phylum Firmicutes)
308
in MAM, whereasin MAM,members of the order Caulobacterales(phylum Proteobacteria)
309
correlated positively with AQP8. These results suggest that AQP8 participates in water
310
absorptionby the microbiota.
311
The microbiota show geographic differences.Although our data for Chinese individualsshowed
312
ahigher
313
correlatedmoststronglywith the parameters and proteins tested. Bacteria can be classed by
314
phylum, class, order, genus, or species, and recent research has predominantlyfocused on the
315
phylum and genus levels. However, in the present study, many genera belonging to the same
316
order displayed the same trends. In examining the functions of bacteria, we must consider the
317
taxonomic level, and the order level may be the best level at which to study bacterial function.
318
The components of MAM and LM were very different, and few correlations were shared by both
319
populations of bacteria. Interestingly, correlations with systemic indicessuch as age, BMI, and
320
BDNF expressionwere mainly observed among the LM bacteria, whereas correlations with
321
systemic and local indices, such as TLR2 expression, were observed in both MAM and LM. The
322
protein ZO1,which usually exerts a local effect, mainly correlated with bacteria in MAM. These
323
results suggest that the functions of bacteria are closely related to their sites of occurrence in the
324
gut. In conclusion, the microbiota of LM and MAM differ. Because microbial population
325
structure reflects function, the different bacteria colonizing different locations in the
326
gastrointestinal tract play different roles by regulating the expression of different proteins.
abundance
ofProteobacteria,
members
of
the
phylum
Firmicutes
References:
327
328
1
Kataoka K. The intestinal microbiota and its role in human health and disease.
329
J Med Invest. 2016;63(1–2):27–37. [PMID: 27040049 DOI: 10.2152/jmi.63.27.]
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
330
2
Ringel Y, Maharshak N, Ringel-Kulka T, Wolber EA, Sartor RB, Carroll IM. High
331
throughput sequencing reveals distinct microbial populations within the mucosal and luminal
332
niches in healthy individuals. Gut Microbes. 2015;6(3):173-181.[ PMID: 25915459
333
DOI:10.1080/19490976]
334
3Sundin J, Rangel I, Fuentes S, Heikamp-de Jong I, Hultgren-Hörnquist E, de VosWM,
335
Brummer RJ. Altered faecal and mucosal microbial composition in post-infectious irritable
336
bowel syndrome patients correlates with mucosallymphocyte phenotypes and psychological
337
distress. Aliment Pharmacol Ther. 2015;41(4):342-351. [PMID:25521822 DOI:
338
10.1111/apt.13055.]
339
4
Piche T. Tight junctions and IBS--the link between epithelial permeability,low-grade
340
inflammation, and symptom generation? Neurogastroenterol Motil. 2014;26(3):296-302.[ PMID:
341
24548256.DOI: 10.1111/nmo.12315]
342
5
Wang P, Chen F, Du C, Li C, Yu Y, Zuo X, Li Y. Increased production of BDNF in
343
colonic epithelial cells induced by fecal supernatants from diarrheic IBS patients. Sci Rep.
344
2015;5:10121.[ PMID:25998025 DOI: 10.1038/srep10121]
345
6
de Kivit S, Tobin MC, Forsyth CB, Keshavarzian A, Landay AL. Regulation of Intestinal
346
Immune Responses through TLR Activation: Implications for Pro- and Prebiotics. Front
347
Immunol. 2014;5:60.[ PMID: 24600450 DOI: 10.3389/fimmu]
348
7
Guttman JA, Samji FN, Li Y, Deng W, Lin A, Finlay BB. Aquaporins contribute to
349
diarrhoea caused by attaching and effacing bacterial pathogens. Cell Microbiol. 2007;9(1):131-
350
141.[ PMID: 16889624]
351
8 Metzler-Zebeli BU, Mann E, Schmitz-Esser S, Wagner M, Ritzmann M, Zebeli Q.
352
Changing dietary Calcium-Phosphorus level and cereal source selectively alters abundance of
353
bacteria and metabolites in the upper gastrointestinal tracts of weaned pigs. Appl
354
EnvironMicrobiol. 2013;79(23):7264-7272.[ PMID:24038702 DOI: 10.1128/AEM.02691-13]
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
355
356
9
Ling Z, Liu X, Luo Y, Yuan L, Nelson KE, Wang Y, Xiang C, Li L.
357
Pyrosequencinganalysis of the human microbiota of healthy Chinese undergraduates. BMC
358
Genomics.2013;10;14:390.[ PMID: 23758874 DOI: 10.1186/1471-2164-14-390]
359
10
360
Yu L, Zhao M, Li Q. Gut microbiota community and its assembly associated with age and diet in
361
chinese centenarians. J Microbiol Biotechnol. 2015;25(8):1195-1204. [ PMID: 25839332 DOI:
362
10.4014/jmb.1410.10014]
363
11
364
differs from that of Americans, Europeans and Asians. BMC Microbiol. 2014;14:311.[ PMID:
365
25495462
366
DOI: 10.1186/s12866-014-0311-6]
367
12
368
receptor 2 pathway establishes colonization by a commensal of the human microbiota. Science.
369
2011;332(6032):974-977.[ PMID: 21512004 DOI: 10.1126/science.1206095]
Wang F, Yu T, Huang G, Cai D, Liang X, Su H, Zhu Z, Li D, Yang Y, Shen P, Mao R,
Escobar JS, Klotz B, Valdes BE, Agudelo GM: The gut microbiota of Colombians
Round JL, Lee SM, Li J, Tran G, Jabri B, Chatila TA, Mazmanian SK. The Toll-Like
370
371
372
373
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
Fig1A
374
375
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
Fig1B
376
377
Fig1. 16S rRNA gene surveys reveal a clear separation of the LM and MAM.
378
(A) Dendrogram obtained with complete linkage hierarchical clustering of the samples from
379
stool and mucosa based on the total OTUs. (B) Principal coordinate analysis (PCoA) plot
380
based on the weighted UniFrac metric.
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
381
382
Fig2. Relative abundances of bacterial phylum in the stool and mucosa samples. Statistically
383
significant differences in all phylum( p < 0.05).
384
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
385
Table1 Quantitative polymerase chain reaction (qPCR) primer used in this study to enumerate
386
specific gene
Primer Sequence 5’-3’
BDNF
ZO-1
TLR2
TLR4
AQP3
AQP8
F
AGGTGGCTCTGGAATGACAT
R
GGCATAAGTCGGCTTGAGTG
F
CAGTGCCTAAAGCTATTCCTGTGA
R
CTATGGAACTCAGCACGCCC
F
TGATGCTGCCATTCTCATTC
R
CGCAGCTCTCAGATTTACCC
F
CAGGGCTTTTCTGAGTCGTC
R
TGAGCAGTCGTGCTGGTATC
F
AGACAGCCCCTTCAGGATTT
R
TCCCTTGCCCTGAATATCTG
F
GGAATATCAGTGGTGGACACT
R
CCAATGAAGCACCTAATGAGC
387
388
Table2 Comparison of the riches and diversity of MAM and LM
MAM
ace
chao
shannon
simpson
160.72±51.62
151.56±48.63
2.29±1.24
0.33±0.32
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
389
LM
195.50±42.04
188.65±42.70
2.901±0.54
0.13±0.07
P
0.004**
0.002**
0.012*
0.001**
**P<0.01 *P<0.05
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
390
Table3 Bacterial taxa that significantly correlated with BDNF
LM
LM
LM
phylum
class
order
familly
genus
Firmicutes
Clostridia*
Clostridiales*
Ruminococcacea
Faecallibacterium*
r=-0.415, p=0.018
r=-0.415,P=0.018
Clostridia*
Clostridiales*
Lachnospiracea*
Lachnospira**
r=-0.415, p=0.018
r=-0.415,P=0.018
R=-0.434,P=0.013
R=-0.588,P=0.00
Bacilli**
Lactobacillal**
Lactobacillace**
Lactobacillus**
r=-0.523, p=0.002
r=-0.524,p=0.002
R=-0.655,P=0.004
R=-0.546,P=0.001
Firmicutes
Firmicutes
R=-0.451,P=0.01
391
392
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
393
Table4 Bacterial taxa that significantly correlated with ZO1
MAM
MAM
MAM
MAM
MAM
MAM
phylum
class
order
familly
genus
Firmicutes
Clostridia*
Clostridiales*
Ruminococcaceae*
Subdohgranulum*
R=0.384,P=0.03
R=0.386,P=0.02
R=0.297,P=0.099
R=0.444,P=0.01
Clostridia*
Clostridiales*
Lachnospiraceae*
Dorea**
R=0.384,P=0.03
R=0.386,P=0.02
R=0.264,P=0.144
R=0.513,P=0.00
Negativicutes*
Selenomonadales*
Veillorellaceae
Megamonas**
R=0.384,P=0.044
R=0.384,P=0.04
Bacteroidia
Bacteroidia
Bacteroidoceae
Bacteroides
R=0.423,P=0.016
R=0.423,P=0.016
R=0.407,P=0.02
P=0.407,R=0.02
Flavobacteriia
Flavobacteriia
Porphyromonadaceae
Parabacteroides
R=0.418,P=0.03
R=0.418,P=0.03
R=0.189,P=0.337
R=0.443,P=0.02
Betaproteobacteria**
Burkholderiales**
Alcaligenaceae**
-
Firmicutes
Firmicutes
Bacteroidetes
Bacteroidetes
Proteobacteria
R=0.456 P=0.00
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
MAM
LM
Proteobacteria
Firmicutes*
R=0.472,P=0.00
R=0.532,P=0.00
R=0.594,P=0.00
Betaproteobacteria**
Burkholderiales**
Oxalobacteraceae*
R=0.472,P=0.00
R=0.532,P=0.00
R=0.553,P=0.05
-
-
-
-
Coriobacteria*
Coriobacteria*
Coriobacteriac*
Collinsella*
R=-0.393,P=0.026
R=-0.421,P=0.029
R=-0.421,P=0.029
R=-0.387,R=0.029
-
R=-0.427 p=0.037
LM
Atinobacteria
394
395
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
396
Table5 Bacterial taxa that significantly correlated with TLR2
LM
LM
LM
MAM
LM
MAM
phylum
class
order
familly
genus
Firmicutes
Clostridia
Clostridiales
Ruminococcaceae
Faecalibacterium
R=-0.522,P=0.00
R=-0.522,P=0.00
R=-0.486,P=0.00
R=-0.593,P=0.00
Clostridia
Clostridiales
Lachnospiraces
Lachnospira
R=-0.522,P=0.00
R=-0.522,P=0.00
R=-0.392,P=0.02
R=-0.400,P=0.02
Bacilli
Lactobacillal
Streptococcaceae
Streptococcus
R=-0.390,P=0.02
R=-0.388,P=0.02
R=-0.466,P=0.00
R=-0.479,P=0.00
Bacilli
Lactobacillal
Streptococcaceae
Streptococcus
R=-0.390,P=0.02
R=-0.367,P=0.04
R=-0.533,P=0.00
R=-0.513,P=0.00
Bacilli
Lactobacillal
Carnobacteriaceae
GranulicatellA
R=-0.390,P=0.02
R=-0.388,P=0.02
Bacilli
Lactobacillal
Firmicutes
Firmicutes
Firmicutes
Firmicutes
Firmicutes
R=-0.373,P=0.03
Carnobacteriaceae
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
GranulicatellA
LM
MAM
MAM
Proteobacteria
Proteobacteria
R=-0.390,P=0.02
R=-0.388,P=0.02
R=-0.696,P=0.00
R=-0.541,P=0.00
Gammaproteobacteria
Pasteurellales
Pasteurellsceae
Haemophilum
R=-0.560,P=0.00
R=-0.560,P=0.00
R=-0.457,P=0.00
Pasteurellales
Pasteurellsceae
Haemophilum
R=-0.650,P=0.00
R=-0.565,P=0.00
R=-0.509,P=0.00
Rickettsioles
Mitochndric
-
Gammaproteobacteria
Proteobacteria
alphaProteobacteria
R=0.605,P=0.03
MAM
MAM
chloroflexi
SAR
Ktedonobacteria
Ktedonobactera
R=-0.696,P=0.01
R=-0.727,P=0.04
Foraminifere
-
-
-
-
-
R=0.522,P=0.00
397
398
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
399
Table6 Bacterial taxa that significantly correlated with TLR4
MAM
MAM
MAM
MAM
MAM
MAM
phylum
class
order
familly
genus
Firmicutes
Clostridia*
Clostridiales*
Ruminococcaceae**
Ruminococcus*
R=-0.371,P=0.03
R=-0.369,P=0.03
R=-0.515,P=0.00
R=-0.442,P=0.03
Clostridia*
Clostridiales*
Ruminococcaceae**
Faecalibacterium**
R=-0.371,P=0.03
R=-0.369,P=0.03
R=-0.515,P=0.00
R=-0.481,P=0.00
Clostridia*
Clostridiales*
Lachnospiraces
Incertae Sedis*
R=-0.371,P=0.03
R=-0.369,P=0.03
Clostridia*
Clostridiales*
R=-0.371,P=0.03
R=-0.369,P=0.03
Clostridia*
Clostridiales*
Peptostreptoco*
R=-0.371,P=0.03
R=-0.369,P=0.03
R=-0.445,P=0.01
Bacilli
Lactobacillal
Streptococcacea*
Firmicutes
Firmicutes
Firmicutes
Firmicutes
Firmicutes
R=-0.370,P=0.03
Lachnospiraces
Blautia*
R=-0.355,P=0.046
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016
-
Streptococcus*
MAM
Verrucomicrobi Verricomicrobiae*
a
MAM
MAM
Proteobacteria
Atinobacteria
Verrucomicrobiales
R=-0.819,P=0.046
Gammaproteobacteria
Atinobacteria
R=-0.383,P=0.04
R=-0.385,P=0.03
Verrucomicrobiales*
-
r=-0.815,p=0.025
Pasteurellales*
Pasteurellsceae**
Haemophilum*
R=-0.446,P=0.03
R=-0.457,P=0.00
R=-0.446,P=0.01
Bifidobacteria*
Bifidobacteriaceae*
Bifidobacteriaceae*
R=-0.485,P=0.01
R=-0.416,P=0.03
R=-0.403,P=0.02
400
401
PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2526v1 | CC BY 4.0 Open Access | rec: 16 Oct 2016, publ: 16 Oct 2016